Recovering from Distributable Thread Failures with Assured Timeliness in Real-Time Distributed Systems